A few changes to the layout test dashboard you might not be aware of if you haven't used it in a while:
- You can see the expected results, actual results and diffs between the two for a given test. This is especially useful when doing rebaseline code reviews. The results shown are the results from the last time the test failed. This includes stacktraces for crashes and text dumps for (most) timeouts. example<http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html#showExpectations=true&tests=LayoutTests%2Fsvg%2Fcustom%2Fmarker-changes.svg%20LayoutTests%2Ffast%2Fdom%2Fobject-embed-plugin-scripting.html> - Supports the webkit.org<http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html#useWebKitCanary=true>and the V8<http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html#useV8Canary=true>canaries - Only shows flaky tests by default. Click the "Show ..." links to show other types of failing tests. - Shows the number of flaky tests per bot in brackets next to the bot name (e.g. Webkit Mac10.5 [16] has 16 flaky tests) - There's a work in progress script for auto-updating test_expectations.txt based off the dashboard. If you are interested in helping out with making this stuff better or have any feature requests, ping me. There are a couple tasks that could use volunteers: - Make rebaseline code reviews in reitveld link to the expectations and current results for tests being rebaselined. This would be really really easy to implement. - Make the layout test dashboard work with upstream's run-webkit-tests (assuming upstream buy-in). This should be easy for someone willing to do some perl hacking. I'm happy to do the non-perl side. :) Ojan -- Chromium Developers mailing list: [email protected] View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev
